Articles/Field notes

Multi-Agent AI

Multi-Agent has a Complexity Problem

AI on the Amstel (HumanX edition) ·Amsterdam, Netherlands

At HumanX Amsterdam’s AI on the Amstel kickoff, industry experts discussed why more AI agents don’t necessarily deliver better results. The panel explored the importance of starting simple, reducing unnecessary complexity, monitoring costs, and improving enterprise AI adoption through effective orchestration and governance.


Multi-Agent AI Has a Complexity Problem

At AI on the Amstel’s HumanX kickoff in Amsterdam, industry practitioners pushed back against the race to build increasingly elaborate agent systems. Their message: start with one agent, prove where it fails, and only then add complexity.

Multi-agent AI sounds compelling.

A network of specialized agents coordinating with each other, handing off work and eventually operating like an autonomous organization.

But at the AI on the Amstel kickoff meetup for HumanX Amsterdam, the discussion was considerably less futuristic — and arguably more useful.

The panel explored what multi-agent AI looks like once you move beyond demos and start deploying it inside real organizations.

The recurring message was simple:

Most companies should probably use fewer agents than they think.

Start with one agent

The panel challenged the assumption that a sophisticated AI system needs to begin with multiple agents.

Instead, the suggested progression was much closer to conventional software engineering:

Start with one agent.

Understand where it succeeds and where it consistently fails.

Then isolate the part of the workflow causing the problem and, if necessary, give that responsibility to another agent.

One example discussed involved an account-management workflow where a single agent initially handled everything: understanding an account, analyzing data and preparing a report.

Only after data analysis emerged as a bottleneck was that responsibility separated into a dedicated sub-agent.

The distinction matters.

The second agent was not added because a multi-agent architecture looked more advanced. It was added because there was a measurable limitation in the simpler system.

A central argument throughout the discussion was that complexity can make systems appear sophisticated without necessarily making them better.

The burden of proof should therefore be reversed.

Do not ask why you should keep the system simple.

Ask why the additional agent needs to exist.

Multi-agent AI is not a token-saving strategy

Another misconception addressed during the session was cost.

Running several agents does not inherently make a workflow more efficient.

In many cases, it does the opposite.

Each agent has to understand its environment, load context, reason about its task, communicate and potentially validate the work of other agents.

That can create enormous duplication.

An example from the discussion illustrated how a main agent might first analyze its environment and formulate a plan, only for a delegated sub-agent to repeat much of the same reasoning.

The optimization did not come from adding another agent.

It came from examining what the agents were actually doing.

That led to one of the most practical recommendations from the discussion:

Read the transcripts.

Agent-to-agent communication can look elegant from the outside while underneath, agents may be repeating context, interrupting each other or performing unnecessary reasoning.

The argument was not that parallel agents are inherently bad.

They can be extremely valuable when speed, specialization or parallel execution justifies the additional compute.

But the metric should be the cost of achieving a useful outcome — including retries, duplicated work and human review — rather than simply counting tokens.

Observability becomes non-negotiable

As agent systems expand, another problem appears:

Understanding what is actually happening.

Once five, ten or potentially hundreds of agents are executing tasks, organizations need to know:

  • What each agent is responsible for
  • Which tools it can access
  • What it costs
  • Whether its outputs meet expectations
  • What happens when something goes wrong

That makes observability less of an infrastructure nice-to-have and more of a prerequisite for deploying agents at scale.

The discussion emphasized the need for evaluation, logging, access controls and clear responsibilities.

This becomes especially important because adding agents can amplify problems rather than solve them.

If the original agent is failing because it lacks context, receives poor instructions or uses an unreliable tool, creating five additional agents can simply reproduce the same failure five times.

The advice was closer to debugging a distributed system than managing an imaginary workforce:

Understand the failure first.

The real AI rollout problem is organizational

Perhaps the most important part of the conversation had very little to do with models.

Building or buying the technology is often the easy part.

Getting an organization to actually use it is harder.

Engineering teams can provide infrastructure, but they do not necessarily understand the detailed workflows of legal, logistics, sales, finance or marketing teams.

Domain experts do.

The result is that successful enterprise AI adoption cannot simply be delegated to a central AI team.

Companies need people inside individual departments who understand both the work and enough of the technology to redesign how that work gets done.

One approach discussed was the emergence of AI architects embedded inside individual teams.

Traditional corporate training was also questioned as a solution.

A mandatory course explaining how to use AI is very different from giving employees access to the technology and letting them discover what it can change in their own workflow.

Managers using the tools themselves, hands-on experimentation and internal power users emerged as more credible mechanisms for adoption.

The underlying message was significant:

AI transformation is increasingly an organizational design problem, not simply a software deployment problem.

Budget: control it — but don't kill experimentation

Here the panel showed more disagreement.

One perspective emphasized establishing budgets and monitoring from the beginning.

Without visibility into AI expenditure, organizations can suddenly discover millions being spent and respond by shutting experimentation down.

Another perspective argued almost the opposite:

Companies sometimes need to accept a temporary "stupid tax" while employees discover valuable use cases.

Premature optimization can prevent those use cases from ever being found.

Both arguments point toward the same tension enterprises now face.

AI experimentation has to be constrained enough that it does not become financially uncontrolled, but open enough that employees are able to discover workflows management could never have designed centrally.

There is unlikely to be a universal answer.

What matters is knowing whether increased spending is producing increased capability.

Don't marry your stack too early

The same caution applied to technology providers.

Models, agent frameworks and orchestration systems are changing too quickly for today's best-performing stack to be assumed to remain the best choice several years from now.

That creates a particular problem for enterprises accustomed to multi-year software agreements.

The discussion explored two possible strategies.

One is to remain deliberately flexible:

Preserve the ability to switch models, mix providers and use specialized agents where they perform best.

The other is to commit deeply to one ecosystem and use the commercial benefits, integrations and discounts that come with that relationship.

While the approaches differed, the underlying concern was the same:

Accidental vendor lock-in can leave companies with neither flexibility nor the full benefits of a deliberate long-term commitment.

Security is still mostly a software problem

When the conversation moved toward AI swarms and security, the discussion challenged some of the more dramatic narratives around thousands of autonomous agents suddenly operating beyond control.

For most companies, the immediate problems are more familiar:

  • Access control
  • Permissions
  • Data exposure
  • Monitoring
  • Tool usage
  • Budget limits
  • Human approval for consequential actions

In other words, much of AI security still resembles traditional software and infrastructure security.

Agents should only receive the permissions they actually need.

Sensitive actions should require additional controls.

External content should be treated carefully.

And organizations need the ability to inspect what happened and stop a run when necessary.

Another concern raised was that the bigger near-term cybersecurity challenge may not be multi-agent coordination itself, but increasingly capable models becoming better at cybersecurity tasks in general.

Small models still matter

During the Q&A, the discussion also turned to smaller and open models.

The argument was straightforward:

Not every task requires the most expensive frontier model.

For narrowly defined work, smaller models can often be faster, cheaper and easier to evaluate.

This becomes particularly interesting in multi-agent systems, where a powerful central agent might delegate clearly defined subtasks to much smaller models.

The principle again comes back to specialization.

If a task is predictable and well-scoped, the most powerful model available may simply be unnecessary.

AI organizations may eventually look like human organizations — or not

Despite spending much of the session questioning unnecessary multi-agent complexity, the discussion was not dismissive of multi-agent AI itself.

Quite the opposite.

Cooperating agent systems could become considerably more important over the next few years.

Companies already divide work between departments, specialists and managers.

It is intuitive that agent systems may eventually replicate parts of that structure:

Specialized agents executing tasks, other agents coordinating them and humans interacting with progressively higher layers of abstraction.

But there was also a counterpoint.

AI systems do not necessarily have to reproduce organizations designed around human limitations.

The eventual architecture may look less like a company made of artificial employees and more like something humans would never have designed in the first place.

That may be the more interesting prediction.

We may end up using increasingly powerful systems while understanding less about why their internal organization works.

The takeaway

The conversation around AI agents has moved very quickly from chatbots to assistants, agents, teams and now swarms.

But terminology can move considerably faster than production systems.

The strongest lesson from AI on the Amstel was therefore not anti-agent or anti-multi-agent.

It was more disciplined:

Start simple. Define the outcome. Evaluate it. Inspect what the system actually does. Add specialization only when you can explain why it is necessary.

A thousand agents make a compelling demo.

One agent reliably solving a real business problem may still be more valuable.


Event: AI on the Amstel — HumanX Amsterdam Kickoff

Location: Amsterdam, Netherlands

Date: September 2026

About the author

Julio Cordal
Julio Cordal

Founder | European Startup Media & Network | AI, Founders & Tech Events

Im the founder of Nectora, a European media company focused on business, technology, entrepreneurship and innovation. We believe the most interesting stories come from the people building the future. At Nectora, we sit down with founders, CEOs, investors and industry leaders to explore their journeys, the decisions behind their businesses, the challenges they face and the ideas shaping tomorrow. Through podcasts, in-depth interviews and short-form content, we bring these conversations to a wider international audience.

More by Julio →

People involved

Grant Easterbook

Moderator · AI On the Amstel

LinkedIn

Anthony Diaz

Panelist · Orq.ai

LinkedIn

Sean Kenny

Panelist · Firmament

LinkedIn

Vladislav Tankow

Panelist · JetBrains

LinkedIn